Cognition, the company behind Devin, raised $2 billion at a $48 billion valuation led by a16z, Accel, Founders Fund, General Catalyst, and Avenir. Annualized run-rate revenue grew from $492M in May to about $900M.
Launched Oct 23, the 2.8T total/104B active multimodal model is live on Kimi.ai, Kimi Code, and the API ( kimi k3.1 pro ), plus a Highspeed variant that runs 2–3x faster at about 2x the Kimi Code quota. In Kimi Code it needs a Plus plan (256K context) or Pro for the full 1M, and thinking can't be turned off (Kimi Code models)…
Released Oct 23, the Monitor tool and plugin monitors now work on Bedrock, Vertex, and Foundry and with telemetry off, and Enterprise custom roles can gate cloud sessions and Remote Control per role. It also fixes Windows PreToolUse and PermissionRequest hooks letting tool calls through when Git Bash is missing and the hook command is only quoted strings (changelog)…
Released Oct 22, it lifts the old mcp<2 pin and migrates MCP transports, errors, schemas, and client lifecycle handling to the 2.x SDK, so code that touches those APIs may need updates. Other breaking changes: AG UI endpoints now need an explicit authorization scope or single user declaration, and the Docker shell's extra run args is renamed unsafe extra run args…
Released Oct 23, ollama run nemotron 3.5 bolt:e4b pulls NVIDIA's Gemma 4 based vision language model for fast local chat, image understanding, and tool use, with a 128K token context window. It ships with a DFlash draft model for speculative decoding to speed up generation…
Released Oct 22, ollama run phonon "..." speaks text aloud as it's generated (or pipe a chat model's reply into it), and apps can stream WAV or PCM audio from /v1/audio/speech using OpenAI's client. It runs on Apple Silicon only for now, English only, with a 4,096 character limit per request (docs)…
Spotted something we missed? Start a thread.